Structure--function relationships of NAD+- and NADP+-dependent dehydrogenases with particular reference to the three-dimensional structure of 6-phosphogluconate dehydrogenase.

subunits A and A' only are modified, this may, of course, be due to an inherent difference in conformation between the A and B subunits or to a conformational change of the B subunits transmitted through the p and/or the r interfaces as resulted from the carboxymethylation of the A subunits. It should be noted that the introduction of a charged group at one of the A subunits has not affected, through the q interface, the carboxymethylation of the other A subunit. It therefore seems unlikely that the removal of a charged group through the decarboxylation reaction from one of the A subunits would affect, also through the r interface, the decarboxylation reaction in the other A subunit. Since, as mentioned above, these photochemical reactions are very sensitive to conformational changes in the enzyme molecule, and no fluorescent derivative is formed when free carboxymethylcysteine and NAD? are u.v.-irradiated even at high concentrations, it would be reasonable to assume that the correct juxtapositions of the carboxymethyl group, the nicotinamide moity of the NAD+ molecule and probably some other side-chain group or groups are essential for the photochemical reaction to take place. The difference in the stoichiometry of the photochemical reactions leading to the formation of the fluorophore and the subsequent decarboxylation for the rabbit muscle, B. stearothermophilus and the tetrakisand bis-carboxymethylated yeast enzymes would seem to suggest some difference in the spatial arrangements of the subunits of the respective enzymes. For the Bacillus enzyme, all four subunits are identical, not only in amino acid sequence but also in conformation. The muscie enzyme might exist, as suggested by some authors (MacQuarrie & Bernhard, 1971; Moras et al., 1975) as a dimer of dimers in which the A and B subunits are somewhat different in conformation (Fig. 7). As for the yeast enzyme, the situation is still more complicated. The tetrakiscarboxymethylated enzyme behaves similarly to the muscle enzyme, whereas the results obtained with the biscarboxymethylated enzyme seem to suggest that even the two A subunits (Fig. 7) may well be slightly different from each other, as are the two B subunits. The difference in behaviour between the tetrakisand bis-carboxymethylated enzymes may have been caused by the fact that the biscarboxymethylated enzyme itself is already asymmetrical chemically. The role of the asymmetrical arrangement of the subunits in some proteins is as yet unclear. In this respect, it should be recalled that the two monomers forming the insulin dimer are also different in conformation (Blundell et al., 1971; Peking Insulin Structure Research Group, 1974). It is possible that such an asymmetrical arrangement is essential to stabilize the oligomeric structure of the protein molecule.